In 2015, the industrial SME Aisa, based in Vouvry, began incorporating AI technology into the machines it manufactures that produce different types of plastic tubing. Machinery manufactured in Valais thus remains at the cutting edge of technology. AI offers more functionality to the end customer and also represents better value for money.
This move into AI has given rise to a whole series of practical projects. The first of these, completed in less than a year, made it possible for the machinery to adjust itself automatically when the production type changes. “By the end of the project, the machine was able to self-calibrate automatically in just two minutes. Before we introduced AI, we had to assign an operator to the task, which would take up to an hour.” The second project involved introducing an automatic quality control system that included cameras, sensors and AI algorithms to check the sizing and external appearance of the tubes produced by the machinery. Other projects in the pipeline include the auto-adjustment of machinery, which involves taking into account a whole host of different parameters. “Given the mass of relevant information available, the only way to manage it all properly is to use AI.”
The budget for this type of project is not excessive. “The cost is mainly labour, because the main task is to identify which parameters to measure and then develop the appropriate algorithms. The fastest projects are completed in less than a year,” says managing director Hugues-Vincent Roy. The machines that Aisa is developing in Vouvry are, in effect, robots.
Identifying hidden potential
The decision to introduce AI at Aisa was initially the result of internal discussions and informal conversations between an engineer working at Aisa and a research scientist at the Idiap Research Institute in Martigny. “Obviously, you can’t do everything in-house; you have to be open-minded and be prepared to talk to other people too.”
Assigning a dedicated team and a Master’s student in AI to the projects gave Aisa a better understanding of the possible applications of AI and of the kinds of skills that were required. “AI is a vast field and there is a lot of hidden potential in a company, especially when it comes to optimising production. At the start, you don’t really know what is possible and what you actually need. So, when approaching the topic, it’s important to be aware of your limitations; you need to discuss the subject with specialists before investing and/or committing dedicated resources.”
Aisa’s adoption of AI has brought about a change of mentality in the company. “It has forced us to go back to the drawing board and look at all the data, abandoning any prior assumptions we may have had. Barriers come down and things that were impossible suddenly become possible. Now that we have a few successful projects under our belts, the mindset within the company has changed, and that’s very positive,” concludes Hugues-Vincent Roy.
